INDEX

Explanations

terms relating to highlighted or emphasized text

oai_token-act-pair · gpt-3.5-turbo

instances of an unusual special character or symbol

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orts

-0.70

 level

-0.66

eking

-0.63

erest

-0.62

ji

-0.62

zel

-0.61

zhou

-0.60

ega

-0.59

pei

-0.58

iple

-0.58

POSITIVE LOGITS

 namely

1.23

perhaps

1.22

albeit

1.11

————

1.02

especially

1.01

particularly

1.01

something

0.98

––

0.96

maybe

0.95

————————

0.95

Activations Density 0.164%

terms relating to highlighted or emphasized text

instances of an unusual special character or symbol

No Comments

No Known Activations